A Parallel AdaBoost Method for Device-Free Indoor Localization

Device-free indoor localization methods based on Channel State Information (CSI) have become an increasingly important technique. The Naive Bayes (NB) classifier has been applied to indoor localization schemes for its simplicity and effectiveness. However, the NB classifier is a weak classifier, leading to unsatisfactory classification accuracy. In order to address this problem, in this paper, we propose the parallel AdaBoost localization method that combines multiple NB classifiers into two strong classifiers based on amplitude and phase. Then, we obtain the weighting coefficient of the classifiers by using the learning strategy, and effectively fuse the results of the classifiers for improving the indoor localization accuracy. Experimental results show that compared with the classical FIFS, CSI-MIMO, PCNB, BLS, ABPS, LSTM and FapFi algorithms, the proposed algorithm has higher localization accuracy.